For a considerable time, silent or synonymous mutations (mutations that do not change the amino acid being coded by DNA) were assumed to be neutral with respect to natural selection. Recent analyses of a variety of organisms indicate that this assumption is not always correct. Almost certainly due to the fact that synonymous codons confer differences in efficiency of translation of messages, selection can detect synonymous mutations. However, both theoretical and empirical studies indicate that the level of selection is very weak, probably on the border of when drift or selection predominates in determining the dynamics. Thus, most studies of such slightly deleterious mutations, a major feature of molecular evolution, are confined to analytical statistical scrutiny (as opposed to empirical experimental manipulations). The emerging set of complete genome sequences for twelve species of well understood Drosophila provides a heretofore unavailable opportunity to study this important feature of molecular evolution. We focus in this proposal on the factors that determine codon usage, especially non-random or biased codon usage. Studies include, first, defining optimal codons for the twelve species of Drosophila to document how much codon usage has changed during the approximately 40-50 MY of evolution. The relationship of level of codon usage bias to both the level and timing (through development) of expression of genes will be explored, with the hypothesis that genes with higher expression have greater codon usage bias as well as this being related to levels of isoaccepting tRNA pools that sometimes change across developmental stages. This will require analysis of specific amino acids. A phylogenetic perspective will address issues such as the relative role of selection, mutation, and drift in determining changes in codon usage. One hypothesis for codon usage bias asserts that misincorporation of amino acids is a major driving factor for usage of optimal codons;this hypothesis will be tested by analysis of the emerging new complete genomes. Finally, we will examine the role of levels of recombination on level and patterns of molecular variation including codon usage. This is made possible by the fact that genes on one chromosome (the "dot") are nearly non- recombining in most species, but have been incorporated into a normally recombining chromsome in D. willistoni, one of the species being sequenced.